8. GRID in Košice

The ATLAS Computing Challenge starts from following conditions:

Average luminosity 1.7x10³⁴ cm^-2s^-1

Trigger rate (in 2007 $\sim$ 160 Hz)

Physics rate (in 2007 $\sim$ 140 Hz)

Running (in 2007 $\sim$ 50 Equiv. Dayes)

Physics events (in 2007 $\sim$ 0.8x10⁹)

It means, that in 2007 before data processing we will have $\sim$ 0.8x10⁹ event sample -> $\sim$ 1.3 PB/year. The estimation for "reconstructed" events, and Monte-Carlo data -> 10 PB/year ( $\sim$ 3 PB on disc), and needed CPU: 1.6M Specint95 including analysis. It is clear, that CERN alone can hadle only a fraction of these resources. The solution how to escape form this situation is grid.
The ATLAS OO software framework is Athena, which coevolves with the LHCb Gaudi framework. ATLAS is truly intercontinental, and in particular, it is present on both sides of the Atlantic

Opportunity: the practical convergence between US, Canada and European Grid projects will come through the transatlantic applications

Threat: There is an inevitable tendency towards fragmentation/divergence of effort to be resisted

On the Fig. 1 the ATLAS computing timeline is shown.

Figure 1: ATLAS computing timeline

On the Fig.2 the schematic view grid-like model for ATLAS is shown, and

Figure 2: Schematic view grid-like model for ATLAS
on the Fig.3 the schematic view of the centre of Fig. 2 - Tier-0 in CERN is shown.

Figure 3: Tier-0 scaling for ATLAS

The goals od DC2 are described on the Fig. 4.

Figure 4: Data Challenge 2

We are now at DC3 period and we can inform you that Košice has been successful connected to DC3.
On the Fig. 5 the grid starting situation in 2005 in Košice is shown.

Figure 5: Number of ATLAS jobs in Košice during per one month and during per one year - starting from 2005